Skip to content

Allow DP attention with DeepSeek V4.1 vision checkpoints #38804

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
kpham-sgl merged 1 commit into from
Sep 10, 2026
Merged

Allow DP attention with DeepSeek V4.1 vision checkpoints #38804

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 2 additions & 3 deletions python/sglang/srt/models/deepseek_v4.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4054,13 +4054,12 @@ def __init__(
self.vision = None
if config.model_type == "deepseek_v41" and config.vision_n_layers > 0:
if (
get_parallel().attn_dp_size != 1
or get_parallel().attn_cp_size != 1
get_parallel().attn_cp_size != 1
or get_pp_group().world_size != 1
or not get_moe_a2a_backend().is_none()
):
raise ValueError(
"V4.1 vision currently supports TP/EP without DP, CP, PP or MoE A2A"
"V4.1 vision currently supports TP/EP/DP without CP, PP or MoE A2A"
)

args = SimpleNamespace(**vars(config), dim=config.hidden_size)
Expand Down
Loading